An authenticated reversible image-encryption architecture is presented for secure image storage and transmission with post-quantum session establishment. The cipher output remains a valid image object and is reconstructed exactly after successful verification. The design protects the image digest and uses it to regenerate the permutation, mask, substitution, and diffusion state at the receiver. The Module-Lattice-Based Key-Encapsulation Mechanism (ML-KEM) establishes fresh session secrets. A hash-based message authentication code using Secure Hash Algorithm 3 verifies package integrity, while the Module-Lattice-Based Digital Signature Algorithm (ML-DSA) authenticates the sender. All required checks are completed before inverse image processing. Formal reductions establish in-transit integrity, sender authentication, and confidentiality for fresh sessions. Experiments give zero reconstruction error, a structural similarity score of 1.0, entropy of 7.9993 bits, a pixel-change rate of 99.61%, and an average intensity change of 33.46%. All 188 applicable component streams in the National Institute of Standards and Technology (NIST) Special Publication 800-22 randomness evaluation satisfied the stated criteria across 64 sessions. A compiled single-core C implementation encrypts 512 ×512 images in 13.8 ms without sender signatures and 15.3 ms with sender authentication, while exact reconstruction is retained through 4096 ×4096 images.
Keeping track of Global Mean Sea Level (GMSL) over long periods is important for assessing coastal risk. However, deep learning models may produce predictions that exhibit short-term increases and decreases, potentially conflicting with expected long-term physical behavior. In this study, we propose a combined Convolutional Neural Network and Isotonic Regression framework, termed CNN-ISO, for modeling and forecasting annual GMSL changes over the period 1880–2024. The framework uses 11 predictors representing atmospheric, oceanic, and cryospheric conditions. To provide longer temporal sequences, the original 144 annual observations are linearly interpolated into 1728 monthly observations. Model evaluation follows a strict chronological split, with training data from 1880–1979, validation data from 1980–1999, and test data from 2000–2024. All preprocessing transformations and the isotonic mapping are fitted exclusively on the training set to prevent information leakage from later periods. On the unseen test period, CNN-ISO achieves R2=0.997 with a 95% confidence interval of [0.995,0.999], together with an MAE of 0.56 mm and an RMSE of 0.72 mm. For comparison, linear regression achieves R2=0.982. Although the predictive improvement over linear regression is relatively modest, CNN-ISO provides an additional constraint by producing monotonic out-of-sample projections, which is consistent with the assumed long-term behavior of GMSL. Under statistical trend extrapolation to 2050, CNN-ISO estimates a GMSL increase of 184.10 mm relative to the baseline. However, estimates vary substantially across models, with linear regression and LSTM producing 116.60 mm and 299.19 mm, respectively. These results should therefore be interpreted as trend-based statistical extrapolations rather than comprehensive climate projections, as the framework does not explicitly model ice-sheet dynamics, changes in ocean circulation, or future emission pathways. Nevertheless, the proposed approach provides a transparent framework for examining model uncertainty and may support early-stage coastal adaptation planning.
Industrial Internet of Things (IIoT) systems generate data from heterogeneous sources across distributed cyber–physical environments, where anomalous behavior may appear as deviations within individual data streams or as inconsistencies among correlated modalities. Existing multimodal anomaly detection approaches remain limited by their dependence on centralized data aggregation, labeled attack samples, or loosely coupled modality-specific learning, thereby restricting their ability to model cross-modal inconsistencies in label-scarce, non-IID, and modality-partitioned IIoT deployments. To address these limitations, this paper proposes a federated multi-scale contrastive learning framework for semi-supervised multimodal anomaly detection. The proposed method integrates modality-specific adversarial autoencoders with federated optimization to learn compact, prior-regularized latent representations of normal system behavior without sharing raw data. Server-side multi-scale cross-modal alignment combines bidirectional InfoNCE with timestamp-valid negatives and a hardest-negative relative-margin objective to refine synchronized modality-specific representations across multiple temporal contexts. At inference, modality-specific reconstruction deviations are combined with learned cross-modal compatibility to construct a system-level anomaly score, with robust normalization and detection thresholds determined exclusively from normal validation data. Experiments on the SWaT and TON_IoT datasets show that the proposed framework achieves F1-scores of 0.9449 ± 0.0064 and 0.9819 ± 0.0033, with corresponding false-positive rates of 0.0119 ± 0.0035 and 0.0385 ± 0.0005, respectively. Additional evaluations under increased client counts, partial participation, imperfect temporal correspondence, missing observations, and alternative architectural configurations further demonstrate stable performance and support the contribution of multi-scale cross-modal alignment and compatibility-aware inference. The results demonstrate the effectiveness of the proposed data-locality-preserving framework for semi-supervised anomaly detection in heterogeneous multimodal IIoT environments.
Fact-checking has emerged globally to combat misinformation and ensure accuracy of information. In Pakistan, it is very intricate due to politically polarized media landscape. This study explores the fact-checking practices of Pakistani journalists amid rampant spread of fake news. Using Qualitative research approach, the researchers conducted semi-structured interviews of 21 journalists (male/female) from mainstream media (electronic & print) and digital media news outlets selected through purposive sampling. The journalists were investigated about methodologies being employed and challenges being faced in fact-checking process. They were also inquired about the impact of fact-checking on quality of journalism and the influence of source's perspective and context of story during fact-checking. Findings reveal that, despite recognizing fact-checking as a professional and ethical commitment, its effective implementation is hindered by organizational policies, external pressures, the absence of proper mechanisms, time constraints, limited professional development training, and the inaccessibility of authentic data and reliable sources. The research suggests that fact-checking is not just a technical or ethical tool in journalism but a democratic requirement in Pakistan. Hence, adopting this practice profusely will reinforce watchdog role of media, and also help to mitigate polarization in social and political spheres by making people more informed citizens.
Abstract Why do foreign investors exit Central and Eastern Europe (CEE) despite formal alignment with EU institutional standards? We argue that divestment reflects not cyclical volatility but strategic responses to institutional fault lines, the non-linear interaction between de jure legal frameworks and de facto enforcement credibility. Analysing panel data (1990–2024) for 20 CEE economies through a hybrid econometric–machine learning framework (ARDL bounds testing, Dumitrescu–Hurlin causality, Random Forest with SHAP diagnostics), we find that: (1) rule-based governance (corruption control) Granger-predicts divestment (Z = 3.897, p = 0.001), whereas formal rule-of-law indicators show ambiguous effects; (2) macroeconomic instability robustly elevates exit risk (β = 0.277, p = 0.001); and (3) trade openness and human capital amplify divestment only when enforcement credibility is weak. Economic scale dominates predictive power (72.0%) but reflects historical FDI exposure, not causal drivers. Critically, machine learning reveals that divestment tipping points emerge from combinations of high inflation, low corruption control, and high openness, patterns invisible to linear models. These findings reframe capital flight as a rational response to institutional dissonance rather than market failure. For policymakers, the implication is clear: EU cohesion policy should shift from formal harmonisation towards performance-based governance that prioritises verifiable anti-corruption enforcement and macroeconomic credibility.
